Skin Cancer Classification

Project Title

Project Area of Specialization

Computer Science

Project Summary

Since the last decade, Cancer has been a severe disease that people are afraid of. Of all the types of Cancer, Skin cancer has been the most serious among the types of cancer, because our human body has the largest organ that is skin. The American Academy of Dermatology Association states that, in the U.S., more than 9,500 people are diagnosed with skin cancer every day. More than two people die of the disease every hour. More than 5.4 million cases of nonmelanoma skin cancer were treated in over 3.3 million people in the U.S. Skin Cancer being such severe disease requires rapid identification of disease. If not so, then it would be considered just acne or pimple. That’s why using latest trends of Artificial Intelligence e.g., Deep Learning, Image Processing etc. would provide rapid classification result of Skin Cancer unlike traditional methods and by using those trends the overall cost of tests for identification of Skin Cancer can be reduced. This project is about providing a model that is able to detect skin cancer. For the training purpose of this project, deep learning and image processing latest techniques would be used so as to increase the precision and accuracy of model which is the main challenge. I would be applying different algorithms of artificial intelligence through which this trained model will be able to provide specific accuracy of the tested images. Subsequently, this project will be deployed as a web based application that would be provide accessibility to the user and ultimately rapid results within the reach of few seconds. This project would open the doors of the latest computing technologies to be used for the field of medical sciences for further research purposes.

Project Objectives

Accessibility: A common person can have access of application using which he can classify Skin Cancer
Time Saving: Time can be saved due to autonomous application of classification of Skin Cancer
Rapid Results: The proposed idea can provide rapid solutions to the end user as compared to traditional methods
Multiple Choice of Models: While training, various datasets would be used
Deliverable: Provide Web Application as a Deliverable at the end and that can also save images

Project Implementation Method

The project implementation includes:

Find out the best algorithms and classification models that works well for image classification
Find out different Skin Cancer Datasets available
Training best available models on my datasets
Retraining the models on different dataset
Testing model after training
Deploying model
Testing models on real-world data
Analysis of data based on obtained result

Benefits of the Project

This project would be developed by being trained on different dataset
This project would allow end users to choose different model for classification
This project would allow user to save the images in their account and load those images

Technical Details of Final Deliverable

Hardware Interfaces
- None
Software Interfaces
- SQLite Database: SQLite database will be used to store the data of users, because it is lite weight
- Django Framework: To provide web based application
- Django File Handling: To store images
- Pytorch: This would be used to support the development, training of model, and testing of deep learning models.
Communications Interfaces
- This web based application can be accessed using browsers from one’s device